Three-dimensional representation of skin structure

Abstract

The present disclosure generally relates to an automated method and system for generating a three-dimensional (3D) representation of a skin structure of a subject. The method comprises: acquiring a plurality of two-dimensional (2D) cross-sectional images of the skin structure, specifically, using optical coherence tomography (OCT) technique; computing a cost for each 2D cross-sectional image based on a cost function, the cost function comprising an edge-based parameter and a non-edge-based parameter; constructing a 3D graph from the 2D cross-sectional images; and determining a minimum-cost closed set from the 3D graph based on the computed costs for the 2D cross-sectional images, wherein the 3D representation of the skin structure is generated from the minimum-cost closed set.

What is claimed is: 
     
       1. An automated method for generating a three-dimensional (3D) representation of a skin structure of a subject, the method comprising:
 acquiring a plurality of two-dimensional (2D) cross-sectional images of the skin structure, each 2D cross-sectional image comprising a skin surface profile; 
 computing a cost for each 2D cross-sectional image based on a cost function, the cost function comprising an edge-based parameter associated with gradient information of the skin surface profile and a non-edge-based parameter associated with homogeneity information of the 2D cross-sectional image above and below the skin surface profile; 
 constructing a 3D graph from the 2D cross-sectional images; and 
 determining a minimum-cost closed set from the 3D graph based on the computed costs for the 2D cross-sectional images, 
 wherein the 3D representation of the skin structure comprising the skin surface profile is generated from the minimum-cost closed set. 
 
     
     
       2. The method according to  claim 1 , wherein computing the costs for the 2D cross-sectional images comprises computing a cost for each pixel of each 2D cross-sectional image. 
     
     
       3. The method according to  claim 1 , further comprising performing skin topographic analysis on the 3D representation to assess skin roughness of the subject. 
     
     
       4. The method according to  claim 3 , wherein the skin topographic analysis comprises performing a plane rectification process. 
     
     
       5. The method according to  claim 4 , wherein the skin topographic analysis further comprises generating a 2D depth map. 
     
     
       6. The method according to  claim 5 , wherein the skin topographic analysis further comprises computing a set of roughness parameters. 
     
     
       7. The method according to  claim 6 , wherein the roughness parameters are calculated based on a sliding window approach on the 2D depth map. 
     
     
       8. The method according to  claim 6 , wherein the set of roughness parameters comprises amplitude and frequency parameters. 
     
     
       9. A system for generating a three-dimensional (3D) representation of a skin structure of a subject, the system comprising a processor configured for performing operations comprising:
 acquiring a plurality of two-dimensional (2D) cross-sectional images of the skin structure, each 2D cross-sectional images comprising a skin surface profile; 
 computing a cost for each 2D cross-sectional image based on a cost function, the cost function comprising an edge-based parameter associated with gradient information of the skin surface profile and a non-edge-based parameter associated with homogeneity information of the 2D cross-sectional image above and below the skin surface profile; 
 constructing a 3D graph from the 2D cross-sectional images; and 
 determining a minimum-cost closed set from the 3D graph based on the computed costs for the 2D cross-sectional images, 
 wherein the 3D representation of the skin structure comprising the skin surface profile is generated from the minimum-cost closed set. 
 
     
     
       10. The system according to  claim 9 , wherein computing the costs for the 2D cross-sectional images comprises computing a cost for each pixel of each 2D cross-sectional image. 
     
     
       11. The system according to  claim 9 , wherein the non-edge-based parameter is associated with a measure of a dark to bright transition at the skin surface profile. 
     
     
       12. The system according to  claim 9 , the operations further comprising performing a skin topographic analysis on the 3D representation to assess skin roughness of the subject. 
     
     
       13. The method according to  claim 1 , wherein the edge-based parameter comprises an orientation penalty function based on gradient orientation. 
     
     
       14. The method according to  claim 13 , wherein the edge-based parameter further comprises a thresholding function that suppresses pixels where a first image derivative is below a first threshold and a second image derivative is below a second threshold. 
     
     
       15. The method according to  claim 14 , further comprising computing the first and second image derivatives using a Gaussian kernel and a Scharr operator. 
     
     
       16. The method according to  claim 1 , wherein the non-edge-based parameter is associated with a measure of a dark to bright transition at the skin surface profile. 
     
     
       17. The method according to  claim 16 , wherein the non-edge-based parameter is associated with a measure of a number of bright pixels above each pixel. 
     
     
       18. The system according to  claim 9 , wherein the edge-based parameter comprises an orientation penalty function based on gradient orientation. 
     
     
       19. The system according to  claim 18 , wherein the edge-based parameter further comprises a thresholding function that suppresses pixels where a first image derivative is below a first threshold and a second image derivative is below a second threshold. 
     
     
       20. The system according to  claim 11 , wherein the non-edge-based parameter is associated with a measure of a number of bright pixels above each pixel.